1b-agent/OneCAD-Draft
1
0 Fork 0
forked from 1b-agent/opencadstudio-upstream
Code Pull Requests Activity

revert: drop wire segment caps that silently truncated geometry

Browse Source 
Remove MAX_SEGS_PER_WIRE = 100_000 and MAX_TOTAL_SEGS = 3_000_000.
These were stop-gap GPU-memory guards that mutated visible output:
single complex entities were cut mid-geometry, and entire wires past
the 3M total were dropped by arrival order with no spatial awareness.
Either is worse than honestly drawing whatever the file contains —
truncated arcs and missing road segments look like bugs to the user,
not "we ran out of GPU memory."

The matching "Temporary Workarounds" section in
RENDERING_OPTIMIZATION.md goes with them. Per-entity LOD and
viewport culling (Phase 1 + 3.2 in the same doc) remain the right
long-term answer; until they land, prefer correctness over a soft
crash on pathological files.

Co-Authored-By: Claude Opus 4.7 <[email protected]>
This commit is contained in:
HakanandClaude Opus 4.7 committed 2026-05-13 10:11:54 +03:00
1 parent 2f751c08f4
commit 9b6d5437fa
3 files changed
+3 -94
Download Patch File Download Diff File Expand all files Collapse all files

No files matched your search

src/scene/pipeline/mod.rs
+1 -30
View File
@@ -600,36 +600,7 @@ impl Pipeline {
}
pub fn upload_wires(&mut self, device: &wgpu::Device, wires: &[WireModel]) {
// Total segment budget: prevents a large file from exhausting GPU memory.
// Each segment → 6 GPU vertices × 96 bytes = 576 bytes; 3M segs ≈ 1.6 GB.
const MAX_TOTAL_SEGS: usize = 3_000_000;
let mut total_segs = 0usize;
let mut skipped = 0usize;
self.gpu_wires = wires
.iter()
.filter(|w| {
let segs = w
.points
.len()
.saturating_sub(1)
.min(wire_gpu::MAX_SEGS_PER_WIRE);
if total_segs + segs > MAX_TOTAL_SEGS {
skipped += 1;
false
} else {
total_segs += segs;
true
}
})
.map(|w| WireGpu::new(device, w))
.collect();
if skipped > 0 {
eprintln!(
"[H7CAD] upload_wires: skipped {skipped} wire(s) — total segment budget ({MAX_TOTAL_SEGS}) exceeded"
);
}
self.gpu_wires = wires.iter().map(|w| WireGpu::new(device, w)).collect();
// Full-brightness copies of selected wires, drawn on top of everything
// in the selection overlay pass so they're always visible.
src/scene/pipeline/wire_gpu.rs
+2 -20
View File
@@ -27,10 +27,6 @@ use crate::scene::wire_model::WireModel;
use iced::wgpu;
use iced::wgpu::util::DeviceExt;
/// Hard cap on segments per wire entity. Entities with more segments are
/// silently truncated to prevent a single complex object from exhausting GPU memory.
pub const MAX_SEGS_PER_WIRE: usize = 100_000;
// ── Vertex layout ─────────────────────────────────────────────────────────
#[repr(C)]
@@ -160,14 +156,7 @@ impl WireGpu {
];
let half_width = wire.line_weight_px * 0.5;
let n = wire.points.len();
let raw_segs = n.saturating_sub(1);
let capped_segs = raw_segs.min(MAX_SEGS_PER_WIRE);
if raw_segs > MAX_SEGS_PER_WIRE {
eprintln!(
"[H7CAD] wire '{}': {} segments exceeds limit ({MAX_SEGS_PER_WIRE}), truncating",
wire.name, raw_segs
);
}
let capped_segs = n.saturating_sub(1);
// Cumulative arc-length per point.
let pts_len = capped_segs + 1;
@@ -270,14 +259,7 @@ impl WireGpu {
let half_width = wire.line_weight_px * 0.5;
let n = wire.points.len();
let raw_seg_count = if n >= 2 { n - 1 } else { 0 };
let seg_count = raw_seg_count.min(MAX_SEGS_PER_WIRE);
if raw_seg_count > MAX_SEGS_PER_WIRE {
eprintln!(
"[H7CAD] wire '{}': {} segments exceeds limit ({MAX_SEGS_PER_WIRE}), truncating",
wire.name, raw_seg_count
);
}
let seg_count = if n >= 2 { n - 1 } else { 0 };
let mut vertices: Vec<WireVertex> = Vec::with_capacity(seg_count * 6);
// Precompute cumulative arc-length at each point.